Hidden semi-Markov model

This disparity has been known for some time, and several attempts were made in ASR to correct this, the most notable proposal being the hidden semi-Markov model (HSMM) [282]. In this model, the transition probabilities are replaced by an explicit Gaussian duration model. It is now known that this increase in durational accuracy does not in fact improve speech recognition to any significant degree, most probably because duration itself is not a significant factor in discrimination. In synthesis, however, modelling the duration accurately is known to be important and for this reason there has been renewed interest in hidden semi-Markov models [504], [514],... 

Zen, H., Tokuda, K., Masuko, T., Kobayashi, T., and Kitamura, T. Hidden semi-markov model based speech synthesis. In Proceedings of the 8th International Conference on Spoken Language Processing, Interspeech 2004 (2004). 

Future research will be focused on the extension of the MB-HMM to a multi-branch Hidden semi-Markov Model (MB-HSMM) due to the fact that the state sojourn time of an HMM follows an exponential distribution, which may not be hold in practice. Another possible extension of the model is the ability of transition between the states of the different branches. The proposed model and the extension should also be validated on the data of real systems. 

Dong, M. D. He (2007). A segmental hidden semi-markov model (hsmm)-based diagnostics and prognostics framework and methodology. Mechanical Systems and Signal Processing 21(5), 2248—2266. 

Peng, Y. M. Dong (2011). A prognosis method using age-dependent hidden semi-markov model for equipment health prediction. Mechanical Systems and Signal Processing 25( ), 237 252. 

U, X, 2013. Bering performance degradation assessment based on continuous hidden semi-Markov model. South China University of Technology, Guangzhou. 

PFAM is a database of Hidden Markov Models of protein families and domains, maintained at the Sanger Centre in Cambridge1651. The concept of PFAM is comparable to that of the PROSITE profile section. Similar to the profiles, the HMMs in PFAM have been derived by the iterative refinement procedure mentioned in Sect. 5.2.4. Unlike the PROSITE profiles, which all have been created manually by the curators, the HMMs in PFAM are generated semi-automatically, which accounts for a slightly lower sensitivity. However, this lack is more than compensated for by the facilitated update procedure, allowing the database to grow much faster than PROSITE and to have a shorter generation cycle. Currently, PFAM holds 2727 entries. 

---